Valorizing industrial side streams through microalgae cultivation: A roadmap for process scale-up

> TRL5). The aim of the present review is to provide a roadmap into microalgae use regarding biomass productivity, biomass concentration, removal of nitrogen, phosphorus and Chemical Oxygen Demand and the percentages of lipids, proteins and carbohydrates reported in the literature to better guide academics and industrial stakeholders on making decision to further optimize and valorize their side-streams.


Introduction
The disproportionate growth in several industrial sectors is a current issue known to produce large volumes of effluents that play a major role in water pollution [1][2][3].In a recent report, data from the European Statistical Office (Eurostat), the Organization for Economic Cooperation and Development (OECD) and the United Nations Statistics Division (UNSD) displayed that the volume of industrial wastewater in 2015 for 32 countries, amounted to 45,311 million m 3 [3].In the case of European Union (EU) Member States, 5293 million m 3 of industrial wastewater were generated by the 16 reporting countries [3].Among the countries that reported both industrial wastewater volume and its treatment, just 30 % of it was treated and only 3 % of these countries subjected their effluents to at least a secondary treatment [3].
Some of the industries with the highest contribution to the generation of waste streams are the agro-industrial, food, and beverage [3] which produce wastewater rich in nutrients [4,108,123].The surplus said nutrients, such as Nitrogen (N), Phosphorous (P), and Chemical Oxygen Demand (COD), if not correctly discharged, can pollute water bodies and lead to a condition known as cultural eutrophication [5,6].Eutrophication has been recognized as the most prevalent water quality issue, reducing the freshwater and marine ecosystems biodiversity [130,131].Recent surveys state that 28 % of freshwater bodies, at a global scale, presented "harmful" qualities and that 65 % of Europe's Atlantic coastal waters exhibit symptoms of eutrophication [7,8].These statistics signal clear failures in the systems in place for wastewater management and remediation.
There are a variety of ways in which industrial wastewater can be minimized and treated in order to reduce pollutant accumulation in the environment [9,10,109].Conventional means for wastewater remediation, such as electrochemical methods, membrane technologies, and precipitation have been proven not to be completely sustainable nor efficient [11].Hence, new strategies have been adopted to reduce the elevated energy and resource requirements wastewater treatment entails [105].In addition, valorization of the waste can lead, eventually, to smaller waste volumes and additional income sources [12,112].
Recently, there has been an increase in research on how microalgae can be utilized to treat wastewaters; N, P (alongside with other nutrients) and COD are reduced by these organisms resulting in what is considered a wastewater tertiary treatment [13,14,116].Microalgae are unicellular photosynthetic microorganisms, and their growth is dependent on the media conditions in which they are cultivated, such as temperature, pH and nutrient content; considering its nutrient load and the feasibility of modifying its parameters according to specific microbial growth conditions (such as pH and temperature ranges), industrial wastewaters provide a perfect environment for them to grow [4,15,16].The main advantages presented by the aforementioned treatment, both over traditional water treatment methods and over the use of other organisms utilized for bioremediation, are their great adaptability to harsh conditions, their resistance to toxic or hazardous materials and the relatively low associated costs when contrasted against chemical and physical treatment methods [17][18][19]106,[124][125][126].Additionally, microalgae are widely known for accumulating considerable amounts of lipids making them of great interest for the biofuel industry [118,127] and generating high value proteins, which are also relevant in the nutraceutical and alimentary markets [20].Moreover, microalgae are sources of valuable phenolic compounds, including the biosynthesis of phenolic acids and flavonoids [21], as well as photosynthetic pigments such as carotenoids and chlorophylls [22].In turn, these attributes turn them into an extremely attractive platform for bioremediation purposes.
There are four main mechanisms by which microalgae accomplish bioremediation: biosorption, bioaccumulation, intracellular biodegradation, and extracellular biodegradation; each of these processes target specific groups of compounds [23].Biosorption occurs mainly as a function of the composition of the microalgal cell wall, which is mostly negatively charged and acts as a binding site for heavy metals and other positively charged emerging contaminants [24,119].On the other hand, bioaccumulation is a metabolically active process by which contaminants are taken into the cell alongside the nutrients required for cellular growth, allowing for additional binding both inside and outside the cell [25].Intracellular biodegradation is arguably the most relevant microalgal mechanism for bioremediation; microalgae can catalytically break down complex organic molecules in the medium and utilize them for their nutrition [26,113].Lastly, extracellular degradation is possible because microalgae secret polymeric substances, forming a biofilm that acts as an external digestive system mediated by the production of extracellular enzymes [27,28].Some of the most promising microalgae and cyanobacteria genera for nutrient removal and water treatment are Chlorella, Arthrospira (Spirulina), and Scenedesmus [29,107].For instance, Chlorella vulgaris and Arthrospira maxima or Limnospira maxima (Spirulina maxima) have been shown to have removal efficiencies ranging from 60 to 100 % for nitrogen, from 65 to 88 % for phosphorous, and from 54 to 86 % of COD [30][31][32][33].These results are both on par with and, in some cases, surpass those of other common water treatment methods [34,35].As for highvalue products, the genera most recognized for accumulating high quantities of lipids are Dunaliella and Chlorella [122], but the first one has not been widely studied in terms of industrial waste stream usage [36].In terms of protein production Chlorella and Spirulina tend to have the highest content, even though there is still a need for cost decreasing and better process scale-up to allow for their mass production [37,101,110,121].
In addition to the foregoing genera, other microalgae strains have been investigated for various purposes.In studies focusing on lipid production and nutrient removal, Arthrospira máxima has exhibited promising results [32].Furthermore, Ascochloris sp. has been studied for its potential in pigment production and nutrient removal [38].Aurantiochytrium mangrovei has been investigated for its capacity to produce carbohydrates and remove nutrients [39].Additionally, Parachlorella kessleri has been examined for its biomass production and nutrient removal efficiency [40].Other microalgae strains such as Chlorococcum and Kirchneriella obesa have also demonstrated potential in biomass production and nutrient removal [40,41].
Since it is hard to maintain axenic conditions and some species have been proven to have great synergy with one another [120], consortiums have been explored to assess their potential [42,111,115].Microalgae consortia enhance nutrient removal and in some cases lipid production in municipal wastewater but there is a need to investigate further regarding the usage of industrial side streams and the improvement of biomass productivity [43,103,114].A summary of the most common types of wastewaters used to obtain specific products and the relevant legislation in the European Union is presented in Table 1.
There is a need for a better understanding of the relationship between the conditions inherent to microalgal wastewater treatment.Hence, in order to ensure the selection of the correct strain, bioreactor type and mode of operation for any given process it is necessary to understand the state of the art when it comes to microalgal remediation and revalorization of industrial wastewaters.Therefore, the purpose of this review article is to collect, present and analyze the reported microalgae production systems that employ industrial wastewater as substrate at a bench scale or higher.Biomass productivity, biomass concentration, removal of nitrogen, phosphorus, and COD, as well as the percentages of lipid, protein, and carbohydrate content, were analyzed considering the type of reactor, its operation mode, operation scale, industrial wastewater type and different microalgae taxon.This review paper clusters the most comprehensive collection of industrial wastewater valorization microalgae cultures trialed in a bench, pilot, or industrial scale.
This review was conducted via ScienceDirect, Google Scholar, and Elsevier between September and November 2022.A total of 41 research articles were selected, analyzed, and contrasted to better elucidate the effect of different parameters in microalgal production and bioremediation performance.A database was developed with the studies gathered, in which the following information was compiled when available: microalgae information (species and strain), reactor information (type, operational volume and operational mode), industrial wastewater characteristics (origin, initial concentration of N, P, COD, pH and temperature of growth), the final product of the study (type of biomolecule, concentration, and use), biomass productivity and final concentration.Each data set was accompanied with information from its source (Digital Object Identifier and APA 7th Edition Cite).The cases compiled were assigned a two-character identifier code (e.g.A0) without any specific criteria.The database can be found in the Supplementary materials.Several search criteria were applied to ensure the relevance of the studies to be analyzed.Firstly, the studies had to have tested the growth of microalgae in industrial wastewater; results for domestic and urban waste streams were excluded because of the perceived difference in composition and nutrient content.Additionally, the studies must have assayed a scale of production of at least 5 L, to ensure that the results were applicable to conditions like those expected in an industrial setting.Lastly, papers that failed to report the maximum biomass productivity and the maximum biomass concentration were discarded because of the lack of comparable data.

Effect of the mode of operation, wastewater origin and microalgae genus on biomass production
Biomass production is an important parameter to measure the cost of a bioprocess.It can be observed a clear tendency, as most of the data points are located between 0.00 and 0.50 g⋅L − 1 ⋅d − 1 of biomass pro-ductivity and 0.00-3.00g⋅L − 1 of biomass concentration (Fig. 2).The ranges for biomass concentration are very variable depending on species and cultivation conditions, such as wastewater composition; it has been observed that they fluctuate between 0.2 and 4.25 g⋅L − 1 [63].Additionally, biomass productivity was equally affected by these factors, and the reported range goes from 0.03 to 0.89 g⋅L − 1 ⋅d − 1 [63].However, some outliers show exceptionally high results in either of the parameters graphed and in some cases in both at the same time.
The most outstanding outlier, in terms of biomass productivity, is the data point E1 [39] which has a value of 3.30 g⋅L − 1 ⋅d − 1 , furthermore, it also presents a higher-than-average concentration of final biomass with 4.60 g⋅L − 1 .E1 has several remarkable oddities, the most telling one being that it is the only case in the database that dealt with the species Aurantiochytrium mangrovei.Hence, the need for further experiments utilizing this species is apparent, since no such additional studies can be found, and it might prove to be a useful alternative to work with.An additional quirk of E1 is the type of wastewater utilized.This case had the advantage of possessing a bountiful carbon source in the form of spent osmotic solution from the candied fruit industry [39] and even though there is not a clear tendency that would suggest that the food production wastewaters result in better outcomes, it is important to not neglect the evident advantage that elevated sugar content represents for algal growth.These elements are, most likely, the ones that made this case highly successful.
Nevertheless, in the best cases of biomass concentration, neither the scale nor the modes of operation are relevant variables to justify elevated values.Higher biomass concentrations seem to be associated to the type of wastewater, for example W0, X0 and O1, (Fig. 2) which have some of the highest reported values, utilize the waste streams of alcoholic beverage production [30,64].Notwithstanding, other cases using this type of wastewater are not as successful; this observation then points to a combination of the microalgae genus and the protocol used for their growth as probable cause.W0 and X0 were experiments made by the same authors with Spirulina sp., they were conducted both in pilot and commercial scale with batch and semi-continuous modes of operation, respectively [30].A likely reason for their outstanding results is that a broad method of strain acclimatization was applied to the microalgae for it to survive the harsh conditions presented by the wastewater.This improved the growth of the culture and increased biomass production to values above 7.00 g⋅L − 1 , even though the productivity was still at the lowest values of the majority.On the other hand, O1 was the only consortium successful in achieving elevated biomass concentrations and had higher productivity than all others.Since several consortiums were also cultivated semi-continuously and/ or on bench scale, there was no suggestion that these variables were responsible for this case's success when compared to other consortiums.Therefore, it can be inferred that the variable that influenced this outcome the most was the type of wastewater, which again indicates that this might be one of the key factors that increase or decrease the productivity and concentration of biomass.All the other consortiums were grown in cattle and crop wastewaters since several strains of microalgae can naturally start growing there, but O1 was grown in a winery's waste stream which made the experiment more controlled, as it had less harsh conditions than other streams and the presence of sugars as additional carbon sources [64].
There are some similarities within the cases that surpassed the mean in either concentration or productivity (Y0, J0, M1, C1, I1 and J1) [65][66][67][68][69] and also with the aforementioned outliers (W0, X0, O1 and E1) [30,39,64].Firstly, most of them were cultivated in batch mode, but some desirable outcomes were obtained from semi-continuous operations, none of the successful cases were carried out in a continuous reactor.This last statement clearly excludes continuous operations from the list of suitable candidates to achieve high productivity and concentration of biomass (Fig. 2.a).Furthermore, the scale of the process does not seem to have a noticeable effect on the reliability of the process, as there is a fair distribution in terms of desirable outcomes independent of if larger or smaller volumes are being used (minimum 5 L to 48,000 L) (Fig. 2.a, .b and .c).Moreover, the type of wastewater utilized is truly diverse when it comes to successful cases; there is not a trend for either better concentration or productivity for any specific wastewater type and none seems to be repeatedly better.Which leads to the conclusion that a good outcome is more dependent on the specifics of each experiment, such as specific conditions or the pretreatments of each wastewater prior to cultivation (Fig. 1.b).Finally, the genus seems to have a high effect on the outcome of the experiment since most of the points corresponding to Chlorella sp. are in general higher in both metrics.
Regarding productivity, it can reach values between 0.86 and 1.00 g⋅ L − 1 ⋅d − 1 (M1, C1) [68,69].There is a higher quantity of data points referring to biomass concentration for Chlorella, compared to other genera, which are higher than the medium with values ranging from 2.80 g⋅L − 1 (S1, Z1) [51,71] up to values of around 4.80 g⋅L − 1 (Y0, J0) [65,66] (Fig. 2.c.).Even if Chlorella showed consistently good results, there were other genera that surpassed its performance, such as the aforementioned genera.Other remarkable cases such as I1 and J1 [67], which are Fig. 1.Graphical representation of two paths of wastewater.Top path, no treatment disposal: About 80 % of industrial wastewater is released into the environment without adequate treatment [70], consequently contaminating and, depending on the composition of the currents, eutrophicating water bodies.Bottom path, wastewater as byproduct or substrate: Microalgae and cyanobacteria culture using wastewater serves as tertiary treatment (nutrient removal) while generating valueadded products such as biofuels, fertilizers, bioplastics, and food supplements.commercial scale semi-continuously operated plants with remarkably high volumes using Scenedesmus sp., and obtained very positive productivity results, of about 1.00 g⋅L − 1 ⋅d − 1 .Even if some cases shared conditions, such as wastewater origin, and similarities in the initial treatment it received; there is a distinction in productivity and biomass concentration obtained.This is due to the size of the reactor and the mode of operation used, as excessively big volumes of wastewater normally obtain small biomass concentrations.Certain treatments, prior to or during cultivation, both to the medium and/or to the microalgae itself, were able to increase productivity.

Effect of the reactor configuration on biomass production
All the cases in Fig. 3.a present a final concentration less than or equal to 2.25 g⋅L − 1 [32] and a biomass productivity less than or equal to 0.21 g⋅L − 1 ⋅d − 1 [38], on average the open systems have the lowest final biomass concentration and productivity, all cases have batch or continuous modes of operation and pilot or commercial scale volume (Fig. 3.a).
Both M0 and C2, the two open systems with the highest final concentration and biomass productivity, were conducted in a commercial scale with batch operation mode; however, 62.5 % of the open systems employed the same operation mode (including the two cases with the lowest concentration and productivity, G0 and G1) thus it cannot be inferred if operating in batches improves or decreases the efficiency in open systems with the collected data (Fig. 3.a) [32,[72][73][74].For tubular reactors, the highest productivity was achieved by C1 and J0, both operating in batch mode, while the lowest productivity was obtained in Z0 and O1 when operating semi-continuously.Although the is a considerable difference in productivity between C1 and Z0 of (0.76 g⋅L − 1 ⋅d − 1 ), it's difficult to attribute these results solely to the mode of operation.Several factors seem to be at play such as production scale, scale-up feasibility and general growth conditions.For example, both C1 and F0 operated in semi-continuous mode at a commercial scale, yet the productivity of F0 is clearly lower (Fig. 3.b).Further exploration of the effect of mode of operation in specific bioreactor types is required to elucidate its actual impact on biomass production [64,66,68,75].E1 is an oddity and should be analyzed independently, so was not included in the figures, since it is an outlier.The biomass concentration achieved for column reactors is considerably lower at a bench scale (in B1, A1, P1, and P1) than at a commercial scale (in S1 and T1) (Fig. 3.c); this is the opposite of what is observed for tubular reactors, where O1 reached 6.10 g⋅L − 1 and C1 2.10 g⋅L − 1 (Fig. 3.b) [60,64,68,76].The highest final concentration (8.11 g⋅L − 1 ) was obtained by Krishnamoorthy et al. [30] when using an improvised reactor (W0 and X0) (Fig. 3.d), although it can be considered an atypical case, the semi-continuous systems presented a considerable final concentration of biomass in all types of reactors except for the tubular ones (Fig. 3.b).Regardless of the type of reactor, the final biomass concentration for the continuous mode of operation was in the range of 0.95 to 2.87g⋅L − 1 , while the productivity ranged from 0.41 to 0.83 g⋅L − 1 ⋅d − 1 in A1, F1 and Z1 [51,60,77,78], showing that this comparison does not show any evident correlation with the operation mode.
Interestingly, the four cases that consume residual water from cattle or crop industries as substrate: G0, F1, A2 and B2, show a trend in which the final concentration increases as a function of maximum productivity, especially clear among the last three cases where a continuous mode of operation is employed [51,74,[77][78][79].Both the initial concentration of Nitrogen (N), Phosphorus (P) and Chemical Oxygen Demand (COD) as well as the species vary in each case except between A2 and B2, where Godos et al. [51] evaluated two different consortia: Achnanthes sp., Nitzschia sp. and Protoderma sp. in A2 and Ankistrodesmus sp., Nitzschia sp. and Oocystis sp. in B2, maintaining the same culture conditions (Fig. 3.a).This makes A2 and B2 particularly interesting by being tailored to compare the synergistic relationships of specific consortia and showing the tangible effect of the selected species.On the other hand, this exception serves to highlight the variability inherent to microalgal biomass production systems, which have drastically different results depending on the different growth conditions proposed in each Although the biomass productivity and concentration appear to be relatively low in the examined cases, the observed coefficients of variation (RSD) of 76.12 % and 62.26 % in both dimensions, and 39.18 % for the nitrogen component (N) across the four cases, highlight the importance of conducting a more detailed investigation into the other common substrate components used in G0, F1, A2, and B2.By employing a sufficient number of replicates, it becomes possible to statistically analyze and evaluate the impact of these components.This rigorous analysis can pave the way for proposing a highly significant optimization of microalgae culture media [80], tailored to specific production objectives.This might be particularly feasible considering that using a commonly generated wastewater they reach their maximum biomass productivity between only 10 and 19 days (depending on each case) and scaling up from bench would not be required.Studies in which a tubular bioreactor was used (Fig. 3.b) presented a higher final biomass concentration and productivity than those that used an open system, except for Z0, since its concentration did not exceed 0.50 g⋅L − 1 and its maximum productivity matched 0.15 g⋅L − 1 ⋅d − 1 .The two highest productivities in tubular reactors were achieved with the batch mode of operation, cultivating Chlorella pyrenoidosa (C1 with 0.99 g⋅L − 1 ⋅d − 1 ) and Chlorella minutissima (J0 with 0.55 g⋅L − 1 ⋅d − 1 ) on a commercial and bench scale, respectively [66,68].On the other hand, with the same type of bioreactor, the highest final concentration of biomass was reached with a consortium of Chlorella vulgaris and Arthrospira platensis (6.1 g⋅L − 1 ) in semi-continuous mode and bench scale [64].The category of reactors with the highest representation in our database was the columntype bioreactor (Fig. 3.c), of which none exceeded a final biomass concentration of 2.73 g⋅L − 1 and productivity of 0.48 g⋅L − 1 ⋅d − 1 [73], reached with Chlorella pyrenoidosa and Scenedesmus acutus separately, except for E1 which, as mentioned above, is considered an outlier.All column reactors used either continuous or batch modes of operation.The lack of representation of studies exploring semi-continuous operation might limit the scope of interpretation of the results.The only case run on pilot scale was U0 with Ascochloris sp.ADW007 operating in batches [38].The effect of the mode of operation is evident, although not drastic, in the column photobioreactor cases A1 and B1 for which Marchão et al. [60] managed to increase productivity from 0.16 to 0.22 g⋅L − 1 ⋅d − 1 by operating continuously and slightly diluting the residual water, which avoided drastically decreasing the final biomass concentration due to substrate inhibition but also decreased enzymatic activity [60]; other variables may have an important effect in this phenomenon, such as the microalgae strain which was Scenedesmus obliquus for both cases.
A considerable number of the data points presented in Fig. 3. d correspond to improvised reactors, which operated in batches and on a bench scale, as is the case for E0, H0, Y0, and X1 [65,73,74,81].Apart from W0, all the cases in Fig. 3.d with a batch mode of operation had a biomass concentration below 5.00 g⋅L − 1 and a productivity of less than 2.76 g⋅L − 1 ⋅d − 1 [65,69].Some studies dealt with and analyzed the effect of up-scaling, such is the case for J1, E0, I0 and X0.None of the cases that increased the reactor capacity showed a subsequent increment of their performance in neither of the response variables (biomass or productivity), except for J1 which, although minorly, improved its biomass productivity [30,38,67,74].Even though the oxygen saturation and working volume was greater in J1 than in I1, other than scale there are no differences between both thin-layer photobioreactors and their operational conditions; however, a greater nutrient removal was achieved by I1 due to its higher surface-to-volume ratio [67].
Ignoring the outliers, the cases that used open systems (Fig. 2.a) and those that used a column-type reactor (Fig. 2.c) are more grouped than the other categories, having a lower range of biomass concentration and productivity.It stands out that the cases within both categories are represented, mostly, by a single type of wastewater (Cattle and crops) and were only evaluated in two modes of operation: batch and semi-continuous.It seems that, by having a smaller variety of variables represented, the results are more homogeneous, which may indicate a possible correlation between the type of reactor, scale and type of industrial wastewater, besides a correct selection of ranges for each variable at least for the open systems and column-type reactors, however more data is needed to accomplish these statements.

Effect of the microalgae genus on biomass production
Biomass productivity and final biomass concentration were analyzed in terms of the genus of microalgae utilized (Fig. 4); this provides a clearer picture of the performance of each of them individually as well as when contrasted with one another.As in all previous images the shape of each data point represents the production scale.Fig. 4.a encompasses the 11 studies making use of Chlorella, with a general range between 0.02 and 1.00 g⋅L − 1 ⋅d − 1 and 0.61 and 4.77 g⋅L − 1 for biomass productivity and final biomass concentration, respectively.With 8 reported cases, consortiums were the second most common choice and had results ranging from 0.01 to 0.24 g⋅L − 1 ⋅d − 1 and 0.40 and 6.10 g⋅L − 1 for biomass productivity and final biomass concentration, respectively (Fig. 4.a).Fig. 4.b, on the other hand, shows the results for the genus Scenedesmus, being the third most common alternative; its ranges are from 0.10 to 1.00 g⋅L − 1 ⋅d − 1 and 0.26 to 1.59 g⋅L − 1 for biomass productivity and final biomass concentration, respectively.Lastly, the genus and species that had 3 or fewer cases each were grouped together (Fig. 4.d), these being: Spirulina, Ascochlorips, Micractinium sp. and Aurantiochytrium mangrovei; most of the results of this graph range from 0.12 to 0.32 g⋅L − 1 ⋅d − 1 and from 0.00 to 2.25 g⋅L − 1 for biomass productivity and final biomass concentration, respectively.It is important to highlight that the main outliers discussed beforehand are present in this figure.When it comes to productivity, the genus Chlorella shows the best results on average, with the lowest average results being those of the studies that make use of consortia.There is no clear effect of the scale of production on the performance of any of the genera, it could be concluded that the results are much more dependent on the growth conditions of each experiment, but these assertions should be further investigated and corroborated by performing statistical analysis in future studies.
A couple of interesting parallels between the two studies with the highest productivity values (M1 and C1) can be observed: both utilized the same strain of Chlorella pyrenoidosa (FACHB-9), the same wastewater with particularly high concentrations of starch acting as an additional carbon source and both included and aeration system that allow for a constant supply of CO 2 to be administered (Fig. 4.a) [68,69].Additionally, one of parameters that might be ensuring these elevated productivities, other than the highly optimized growth conditions, is the high tolerance of Chlorella pyrenoidosa to ammonia which might be inhibiting the growth of other species and genera [82].On the other hand, Y0 and J0 being the points with the overall highest biomass concentration, share almost nothing in common; J0 utilized Chlorella minutissima a saltwater species and Y0 utilized Chlorella vulgaris a freshwater microalga; as could be expected from this mayor difference the wastewaters had starkly different compositions [65,66].Arguably this shows that ensuring the affinity between the algal species and the wastewater is much more relevant than other parameters such as bioreactor type or mode of operation.Fig. 4.b offers a lot of insight, even if the results presented are generally underwhelming, particularly when considering the consensus that microalgal consortia outperform single species cultures [83,84].When analyzing them on a case by case basis, a clearer picture of what might have gone wrong emerges; in most of the studies the consortia were either taken directly from nature (G0, H0, I0, Z0, A2, and B2) [51,74,75,79] or mixed with limited considerations on exploring synergetic conditions of the individual species (O1 and N1) [31,64].None of the studies analyzed the performance of the consortia vs individual algal cultures and in cases such as A2 and B2 direct competition between the algae present could be observed [51,79].O1 stands out as a clear outlier, but this can easily be attributed to the elevated concentration of sugars present in the wastewater utilized, rather than to the effect of the interaction between algal species [64].It is important to clarify that for most studies in this group, the goal was not to achieve elevated biomass productivity, but to ensure nutrient removal; the results for this metric are presented in Fig. 4. The low biomass concentrations presented on Fig. 4.c, could signify that the genus Scenedesmus is not well adapted to proliferate in industrial wastewaters.The main outliers in this case are I1 and J1; both data points come from the same study and therefore share most of their growth conditions.The main differentiator between them and the rest of the studies is the total nitrogen content of the wastewater while all the other studies utilizing Scenedesmus had nitrogen values lower than 50 mg⋅L − 1 J1 and I1 both had an initial concentration of 210 mg⋅L − 1 showing a clear relation between the nitrogen concentration and the biomass productivity achieved [67].
The main outliers mentioned in the previous sections of this study (W0, and X0), are shown in Fig. 4.d (E1 is not included in the graphs).Other than the factors mentioned beforehand, the studies had wastewater highly loaded with sugars acting as an additional carbon source, resulting in mixotrophic growth and explaining, to an extent, the elevated biomass concentration and productivity [30,39].There is a clear need for the diversification of the species studied when it comes to industrial wastewater treatment; as seen during this section the affinity of the microalgae with the wastewater is a highly determining factor for efficiency and therefore exploring additional alternatives might prove bountiful.

Advances of high value-added products from microalgae
Besides the use of wastewater as a culture media, the valorization of the side streams is also a relevant topic to be studied.As mentioned before, some of the products that microalgae can produce include lipids, proteins, and carbohydrates, which give an extra value to their cultivation.Nowadays, more biorefineries are using microalgae strains such as Chlorella, Scenedesmus, Dunaliella, etc. as they can be cultivated in either autotrophic, heterotrophic and mixotrophic conditions and have very minimal nutrient requirements [85,97].

Lipid production
For some of the studies, the generation of a value-added product, besides water treatment and bioremediation, was also contemplated.Developing new products that could be used for consumption or health, utilizing a stream that previously was considered a waste, can be positive both economically and environmentally [102].Most of the cases that reported additional bioproducts dealt with lipids, specifically fatty acids such as omega 3 or other lipids commonly used for biodiesel production (Fig. 5).Additionally, proteins (A1, B1, C1, F0, F1, Z0 and M0) and carbohydrates (C1, M0 and Z0) were also produced in a minority of cases (Figs. 6 and 7).
As previously mentioned, lipids were the most reported biomolecule produced using industrial wastewater for the cultivation of microalgae.Most cases presented a lipid content between 0 and 30 % and tended to have some of the lowest biomass productivity; this is to be expected, since growth is negatively correlated with the amount of lipids accumulated in the cell.This interaction is evident in cases such as G1 and R1, which had the first and second highest lipid contents, respectively, and in turn some of the lowest biomass productivities between the cases mapped in Fig. 5. On the other hand, M1 and C1 which had comparably lower lipid values, possessed the two highest biomass productivities of the cases presented in Fig. 5.The least successful cases were in their majority not focused on value added product generation, but on bioremediation or biomass harvest, such is the case of A1, B1, C1 and T1 [60,68,86].None of these experiments exceeded 10 % lipid content.On the other hand, the best outcomes go from 30 % up to almost 90 %; with most of the cases situated between 30 and 40 %.However, it seems that there is a correlation between successful results and the genus selected for the process.Aschochloris sp.presents consistent results in terms of both lipid content and biomass productivity, slightly outperforming most other species with the exception of Chlorella sp.Its performance seems to be independent of the scale of the operation, which appears to be neglectable.Given that only 4 cases making use of this microalgal species fell inside the parameters of this study, it may be a suitable candidate for more research and further experimentation.Nevertheless, all the articles mentioning this genus used dairy wastewater, which seems to imply that initial nutrient content is highly relevant for lipid production and accumulation [38,73].E1 shows promising results in terms of lipid production because it also maintains a high biomass productivity; however, it is probable that the success of this process is due to the conditions used during experimentation rather than the microalgae species itself or the scale of production [39].Other species that are considered relevant for lipid production are Phaeodactylum sp.,   which have been studied widely.These species have been cultured for lipid production using oil field produced wastewater or dairy manure wastewater as media [132,133].Nevertheless, there have not been experimentations in any large-scale reactor, which excludes them from the scope of the research in which the bioreactors considered are 5 L or more.
When observing the highest results, in terms of lipid content and biomass productivity, which were all obtained with Chlorella sp.R1 and J0 [49,66] they show clear similarities in scale, aim, and protocol.The aim of both experiments was to treat the wastewater and then focus on lipid production.This also explains why the lipid content is not as high as in G1, furthermore, both cases were cultivated within the span of several days while G1 was a months-long procedure.The highest percentage of lipids reported was in the case G1 [72]; having a cellular concentration of lipids of almost 90 %.This procedure was done in a high-scale reactor (raceway pond) and focused exclusively on added value products.However, it was also a 400-day procedure that showed an increase in the amount of lipids after each recovery.Although a high lipid content resulted in a low biomass productivity, the protocol followed to extract said lipids after each month and the medium recycling process improved the percentage of lipids by the end of the experiment [72].While it has a remarkably high lipid content, its lipid productivity and biomass productivity were exceptionally low, which means that there was not a big quantity of microalgae present in the culture, but most of their weight was composed of lipids.
Biomass productivity was plotted against lipid productivity (Fig. 5. b).This comparison gave different results for some cases, but the main conclusion is similar to that drawn from Fig. 5.a, since Chlorella sp.continues to be the most successful.R1 and J0 again have the highest values due to the protocol followed because the objective of the study was mainly biodiesel production.Unlike in lipid content, E1 had the highest lipid productivity of all cases; this means that even though there was not a big amount of lipids inside the cells, the Aurantiochytrium mangrovei has a fast and efficient production of them.
In general, some species are widely known for the production of lipids such as Dunaliella sp. and Nannochloropsis sp., but there was no available information for their use in industrial water.This might be because most of the research done was focused on bioremediation, more common with the Chlorella genus, rather than only value-added product generation.It is also relevant to mention that the characteristics of wastewater tend to be harsh, which is why certain species, which are better adapted to such conditions, manage to outperform those who have shown better results in controlled settings.

Protein production
The second most common bioproduct produced using industrial wastewater for the cultivation of microalgae, considering the cases re-ported in the database, was protein.The highest reported content amounted to 60 % (Fig. 6) in C1, besides having the largest protein percentage, C1 had a biomass productivity of 1.00 g⋅L − 1 ⋅d − 1 , which was the highest of all the tubular photobioreactors.Considering that the average protein dry weight percentage reported in literature for Chlorella pyrenoidosa is lower (57 %) [87], C1 has a potential application for the revalorization of industrial wastes as a microalgal based products [68,88].
Protein content was shown to be inversely proportional to lipid content in all reported cases [32,33,60,68,75,77,78].This can be attributed to the fact that microalgae tend to store much more lipids under stress conditions, such as nutrient deficiency, low or elevated temperatures and high salinity, which also reduce protein content [88,89].
Consistent results are observed in the three cases that make use of Chlorella sp.(C1, F0, and F1).They share none of their growth conditions (nutrient content, type of wastewater or to reach maximum productivity), except for the scale of production, which was commercial (Fig. 6).This information implies a direct relationship between biomass productivity and protein content of the Chlorella genus.With a greater number of reported cases, the effect of each strain (C.pyrenoidosa FACHB-9, C. vulgaris UTEX2714 and C. vulgaris FACHB-8) could be better surmised regarding biomass productivity and protein content.

Carbohydrate production
When compared to lipids and proteins, there is a distinct lack of research focusing on the production of carbohydrates using industrial wastewater as a substrate for microalgae cultivation (Fig. 7).However, the percentage of carbohydrates obtained with a consortium collected from an anaerobic digestion slurry, and conformed by Chlorella pyrenoidosa, Micractinium pusillum, Actinastrum hantzschii, Micractinium sp. and Chlorella colonialis, amounted to 39.65 %.This value is high compared to other production systems that used wastewater as a substrate [32,68,75] but falls short against the 77.60 % reported by Cheng et al. [90] in Chlorella sp.AE10 by adaptive laboratory evolution in BG11 medium [90].The information gathered cannot be used to state if there is a best system to produce more carbohydrates since there is not a clear tendency followed by all the cases and there is a lack of studies where its production is measured.

Nutrient removal efficiencies
Microalgae can also be used to remove nutrients from the water streams.The main reported ones are nitrogen, phosphorus and COD.
The main indicators for bioremediation, being the percentages of removal for total nitrogen, total phosphorus, and COD, were graphed and contrasted (Fig. 8).From all the cases compiled in the database, only 21 included data on the bioremediation indicators.A tendency of elevated efficiency is equally observable in the three metrics; The average removal rate for nitrogen was 71.54 %, it was 74.76 % for phosphorus, and 71.58 % for COD.Given the amount of variability shown in the results of the studies, the nutrient removal efficiencies appear to not have a linear relationship with the biomass productivity performance.This conclusion falls in line with the results obtained by Paddock et al. [91] and by Choi and Lee [92], which point to a complex network of interactions between the microalgae, native microorganisms found in the wastewaters, and growth-inhibiting compounds as the main determining factors when it comes to both biomass production and nutrient removal efficiency.Specifically, members of the genus Chlorella had consistently high removal rates while maintaining elevated productivity [93].On the other hand, the members of the genus Ascochloris have the most consistently positive results in all metrics of removal while maintaining decent levels of productivity.The genus Scenedesmus had the most internal variation in terms of results, both when talking about removal rates and productivity.Determining an ideal balance between high productivity and elevated levels of nutrient extraction would ensure the viability of these processes when implemented in production plants.These results show the amount of adaptability inherent to each strain; with the genus Chlorella, is known to have an elevated tolerance to the inhibitory effects of contaminants such as nitrates and nitrogenous compounds [77,78,128].
It has been shown that mixotrophic growth leads to a higher cell concentration and therefore to a greater bioremediation capacity [94].This is clearly exemplified by C1 being the most efficient data point presented.The one area where C1 seems to fall behind is in the reduction of COD which can be explained by the oversaturation of the wastewater with organic compounds [68,117].C2, E0, and U0 originate from consecutive essays developed by the same research group, utilizing the same wastewater and therefore sharing similar growth conditions; their results show that despite sharing the second highest concentration of COD (7110) they managed a removal rate of approximately 95 %, with similar results with both nitrogen and phosphorus [38,73].Despite maintaining average productivity, the genus Ascochlorips shows great adeptness for bioremediation.The studies dealing with the genus Scenedesmus show a tendency to favor nitrogen removal over phosphorus and COD reduction, as can be seen in the data points Q1, A1, and B1; The exceptions being H1, where phosphorus removal was slightly favored and P1 where both nitrogen and phosphorus were entirely removed, but this might be due to the particularly low content of phosphorus in both cases 4.26 and 1.6 respectively [60,76,95].In all instances, members of the genus Scenedesmus seem to have allowed performance at high COD concentrations, compared to the other genera.Lastly, the consortia show the inverse behavior, being much more adept at bioremediating phosphorus and COD instead of nitrogen with the sole exception being N1, which is a consortium mainly governed by Chlorella vulgaris and therefore explaining its affinity for nitrogen [31,96,98].

Conclusion and further perspectives
The use of industrial side streams as a growth medium for microalgae has been thoroughly explored; limitations on the data available for larger scales remain.Further integration of microalgal bioremediation into the industry and the development of value-added products would increase sustainability and profitability.Based on the data gathered it can be stated that microalgae growth is affected greatly by species, type of wastewater, and processes prior to cultivation.The impact of scale, mode of operation, and bioreactor seems negligible.Finally, the limitations of microalgal production systems can be overcome by scientific advancements, such as genetic engineering, bringing forward a circular economy.

Fig. 2 .
Fig. 2. The present graphs compare biomass concentration (g⋅L − 1 ) with biomass productivity (g⋅L − 1 ⋅d − 1 ), each subfigure categorize the data by type of process (a), microalgae and Cyanobacteria genus (b), and wastewater substrate origin (c).The data marker type categorizes bench (square), pilot (triangle) or commercial (circle) scale.The outlier E1 was omitted in all graphs to better examinate the results, this was replicated in all figures.

Fig. 3 .
Fig. 3. Final biomass concentration versus biomass maximum productivity, expressed in g⋅L − 1 and g⋅L − 1 ⋅d − 1 respectively.Four graphs were made in order to compare the cases depending on the configuration or type of reactor.All cultures in open systems are encompassed in a; tubular reactors in b; column reactors in c; and, the rest of the configurations reported in the literature were plotted in d.Data marker type categorize bench (square), pilot (triangle) or commercial (circle) scale.

Table 1
Products commonly obtained from microalgae by industrial wastewater type and relevant legislation in the European Union for their adoption.